67 Q UA RT E R _ 0 2 _ 2 0 2 6 rency, always on, instantaneous. You want it to be safe, and you don’t want any complexity. This is where tokeni zation in the chain is very valuable, because we can do this within the existing banking system. So that’s tokenization; it’ll be coming to equity markets, it’ll be coming to oil, it’ll be coming to multiple different areas, and it’s nice and secure. - - - - An online version of this interview, featuring audio of the conversation, is available on McKinsey.com EK : There seems to be a prevailing incongruence where many CEOs I speak to insist that the layoffs we are starting to see have nothing to do with AI, while others might suggest otherwise. What are you seeing, and what is your expecta tion of the way AI will change the labor market? JF : What we’ve seen so far is the enormous build- out of AI infrastructure, and that has certainly been helping boost economic growth. But when we look at AI itself, it hasn’t really come through yet. We’re only in year three of this huge AI shift, so I don’t think we should be falling off our chairs that the big productivity benefits haven’t come through yet. What we’re seeing with AI right now is more bottom up, task oriented. At Citi, we’ve given our employees a lot of different AI tools, and they’re augmenting productivity. Our AI usage is high, at about 70 percent, but that’s helping people do their jobs and removing some of the drudgery ele ments. The productivity benefits come from scaled use cases applied top down. We’re systematically going through all our major processes. I use a dishwasher analogy: How you wash a dish by hand and how a dishwasher washes a dish are completely different. How the two pro cesses work has almost nothing to do with each other. We’re looking at AI and applying it to our processes like a dishwasher, as opposed to only thinking about how to better auto mate the manual process. And that is starting to deliver some staggering productivity numbers, particularly with agentic AI. Our coders today are 9 percent more productive in their entirety—that’s 30,000 coders—than they were at the beginning of the year. So I know the agentic piece, where you’re getting agents to act and work together, will be a game changer for some business models. You can see this happening, and you can also see new jobs and the shifting of roles. But what will the timing of this be? Will they happen together? I think we’re in for a bit of a bumpy ride on the pro ductivity operational side. foot and clear about the strategy and direction.’ The other element is the risks. We all talk about what AI means from the operational side, what it’s doing for customer service or coding and so on. From a cyber perspective, as a bank, we have many billions of attacks on us a day—authentication, fraud, money laundering, cyber. It’s a scary combo that keeps me awake at night. So while there are new elements and new threats, there are new jobs there too. It’s going to be a very different world. EK : People talk about your willingness to bring in external talent, which, in my experience, is some thing few CEOs do, at least with great success. How do you think about doing that? JF : We have fantastic talent inside the bank, and sometimes it’s hard bringing external talent into your management team in a company that’s as old as ours. But to my mind, it was an imperative, so we just went about finding the best talent. If you’re trying to build something new, you can often get people who are the best in the world at something, because they’re starting to get bored where they are—they want to build again and demonstrate that they can shape more than one legacy. We were able - - -